Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
December 1, 2007 / Vol. 32, No. 23 / OPTICS LETTERS 3447 Purkinje imaging system to measure anterior segment scattering in the human eye Juan M. Bueno,1,* Dirk De Brouwere,2 Harilaos Ginis,2 Ioannis Sgouros,2 and Pablo Artal1 1 Laboratorio de Óptica, Centro de Investigación en Óptica y Nanofísica, Universidad de Murcia, Campus de Espinardo, 30071 Murcia, Spain 2 Institute of Vision and Optics, University of Crete, Heraklion, Crete, Greece *Corresponding author: [email protected] Received August 28, 2007; accepted October 2, 2007; posted October 29, 2007 (Doc. ID 86991); published November 28, 2007 We present an instrument based on Purkinje imaging that permits the objective measurement of the amount of scattering associated with the eye’s anterior segment, avoiding the contribution from the retina. The experimental system records the fourth Purkinje image, and adequate processing is used to compute a parameter that quantifies the scattering. The method was first tested in an artificial eye and later in normal young eyes wearing customized contact lenses that induced different amounts of scatter. We were able to detect scattering increments, which indicates that this technique may be used as an objective tool to quantify the level of scattering in the anterior segment of the living human eye. The future use of this technique in clinical environments might help to estimate the level of corneal haze in eyes undergoing refractive surgery or/and scattering within the lens during cataract development. © 2007 Optical Society of America OCIS codes: 170.3880, 290.5820, 330.5370, 170.4470. Light scattering is a physical phenomenon inherent to light propagation that is due to the presence of optical inhomogeneities. In the particular case of the eye, beyond aberrations [1], intraocular scattering is a major factor responsible for the decrease in optical performance. This reduces the contrast of the retinal images, and although it is low in young healthy eyes, it becomes more significant with age, corneal surgery, and some pathologies [2,3]. It has been shown that corneal scattering (i.e., haze) [4–6] is one of the negative side effects in corneal refractive surgery. Most techniques used to estimate ocular scattering are subjective, requiring active participation of the subject and a careful understanding of the corresponding tasks [7]. However, some authors have reported objective approaches [8–12]. These provide information on the total ocular scatter, that is, the combined contributions coming from the different ocular components and the retinal reflection. In this Letter we propose a method to separate out the contribution of the anterior segment of the eye (cornea and lens) from that of the retina to the overall scatter. It is based on the analysis of light reflected from the back surface of the lens (fourth Purkinje image [13], PIV) to assess a parameter of scattering (POS) to be used objectively as a tool to quantify different levels of scattering produced in the cornea and/or the lens. Figure 1 shows a schematic of the setup used to image PIV in both artificial and human eyes. A 633 nm collimated He–Ne laser beam passes an artificial pupil (AP) and goes through a removable neutral density filter (RNDF, with an optical density of 1) used to modify the intensity of the incoming beam according to two different experimental conditions. In the registration pathway, a photographic objective conjugates PIV with the plane of a CCD video camera. PIV was slightly defocused on purpose to increase the dy0146-9592/07/233447-3/$15.00 namic range. During image acquisition, subjects were asked to fixate using their nontested eye to a target positioned in a manner to direct the measured eye towards the camera axis. This made the illumination and the recording pathways approximately 40° apart in the horizontal plane and ensured PIV to be distant enough from both the (much brighter) first Purkinje image and the pupil edge (see Fig. 2). The exposure time was the same for all subjects and experimental conditions. Series of PIV were consecutively recorded for two different experimental conditions: with and without the RNDF. For the former condition PIV was not saturated and was thought to capture the directional light. Its intensity was large enough to be well different from light coming from the surroundings. When Fig. 1. (Color online) Schematic of the Purkinje imaging system. AP, aperture to set the size of the beam reaching the eye (this filled the entire pupil). Other abbreviations defined in text. The artificial eye was mounted on translation stages to facilitate its alignment. © 2007 Optical Society of America 3448 OPTICS LETTERS / Vol. 32, No. 23 / December 1, 2007 Fig. 2. Registered image for the human living eye showing the first (circled) and the fourth (squared) Purkinje images. The image size is 4.5 mm⫻ 4.5 mm. the RNDF was removed from the illumination pathway, PIV was saturated. This enhanced the information of the tails of PIV, where the scattered portion of the light is located. All calculations were done using a purposely developed MATLAB (Mathworks, Inc.) script. For each acquired image a square area centered on PIV was extracted (Fig. 2). The averaged intensity radial profile centered on this Purkinje image was computed and normalized to 1. Once the radial profiles are known, the areas under the profiles are computed and named as k⬘ គ nonsat and k⬘ គ sat for the nonsaturated and saturated images, respectively. In particular, k⬘ គ nonsat corresponded to the area under the radial profile of the image between 0 and 0.24 mm. For k⬘ គ sat the range was the same, but excluding the saturated central pixels (i.e., those with a value of 1 in the normalized profile). POS was finally computed as POS = k⬘ គ sat k⬘ គ nonsat . Fig. 3. Scatter-customized contact lenses (left, CL#1; right, CL#5). Both have the same power 共0 D兲, radius 共7.8 mm兲, diameter 共9.2 mm兲, and thickness 共0.225 mm兲. The concentration of microspheres responsible for the amount of scatter introduced was 0 (CL#1) and 0.00109 g / ml (CL#5). Eight normal young eyes from different subjects were involved in the study and used as a control group. In a subgroup of 2 eyes (#7 and #8) the complete procedure was carried out. That is, nonsaturated and saturated PIV were acquired for those eyes wearing the two CLs. Since the unique difference is the amount of scatter introduced at the corneal plane by the CLs, this will allow us to determine that any significant change detected in the light intensity distribution of the image is due to the isolated effect of a variation in light scattering. As an example Fig. 4 shows four images corresponding to the two experimental conditions and the two CLs. In Fig. 5 the values of POS for the 8 (naked) control eyes are depicted. We have also included the POS for the artificial eye with CL#1 as a reference. This plot shows that in normal control young eyes, the amount of scatter (measured as POS) at the anterior chamber is similar for all subjects. Finally, in Fig. 6 we have plotted the results for subjects #7 and #8 and the artificial eye. As a comparison we have also included the mean POS for the control eyes in Fig. 5. Results 共1兲 To estimate the sensitivity of our instrument, the procedure was first tested in an artificial eye (see Fig. 1). This consisted of a scatter-customized contact lens (CL) acting as the cornea, placed in front of a polymethyl methacrylate biconvex intraocular lens (IOL), acting as the lens. CLs are hard [rigid gas permeable (RGP)] lenses (null power) with a well-defined concentration of glass microspheres suspended in the RGP material and were manufactured by Menicon Ltd. (Nagoya, Japan) for the simulation of corneal scatter. The microspheres have a diameter ranging from 1 to 20 m to produce a narrow-angle forward scatter distribution similar to that of the cornea caused by relatively large structures such as keratocytes and protein depositions [14,15]. Figure 3 shows a dark field photo of the CLs (CL#1 and CL#5) used here, where differences in the scattering produced can be observed. Fig. 4. Images of the fourth Purkinje image in subject #7 for nonsaturated (left) and saturated (right) conditions with both CLs. Each image subtends 0.8 mm. December 1, 2007 / Vol. 32, No. 23 / OPTICS LETTERS Fig. 5. POS values for 8 control naked eyes and the artificial eye with CL#1. 3449 the intensity at the halos increases with the amount of induced scattering. The computed POS was sensitive enough to discriminate different induced levels of scattering at the anterior chamber. In older eyes this technique may be an alternative to the current clinical subjective observations of changes of transparency in the ocular media, mainly associated with cataracts [16,17]. The implementation of this technique in a clinical environment might help to quantify the amount of corneal haze following refractive surgery. This also will be useful in detecting an increase of scatter in (early) cataract eyes. An appropriate combination of this instrument with others providing the amount of total scatter could help in the diagnosis of some retinal pathologies associated with an unusual increase of retinal scattering. This work was supported by the European Research Training Network “Sharp-Eye” and the Spanish Ministerio de Educación y Ciencia (grants FIS2004-2153 and FIS2007-64765). References Fig. 6. Values of POS for subjects #7, #8, and the artificial eye, wearing the two scatter-customized CLs. The average POS for the (naked) control eyes computed from data in Fig. 5 has also been included as a comparison. show that in eyes wearing CL#5 the values of POS increase significantly: 16% and 23% for subjects #7 and #8, respectively. In these experimental conditions, the amount of corneal haze is the only difference, which means these changes found in the measured POS are strictly due to the contribution of this induced scatter. For all eyes and within a CL, the values of k⬘ គ sat were always larger that those of k⬘ គ nonsat. However, when comparing both CLs, the values of k⬘ for nonsaturated images were similar. On the other hand, for saturated images there are noticeable differences between both CLs. In particular, for the artificial eye k⬘ គ sat for CL#5 was 34% higher than CL#1’s. The halo around PIV contains contributions of the scattering due to roughness of the posterior lens surface and the bulk scattering in both the cornea and the crystalline lens. Light scattered from structures near the PIV plane will contribute less to the peripheral halo around the image, as the scattered light will be almost in focus, coinciding with the directional image. For this reason, light scattered in the cornea results in a more peripheral halo than light scattered in the crystalline lens. In summary, we have reported a technique to objectively extract the contribution of the ocular scattering coming from the cornea and the lens. This is based on a Purkinje imaging instrument. The procedure has successfully been tested in an artificial eye and in normal healthy eyes wearing scattercustomized CLs. Results show that the core of the image hardly changes when increasing scattering, but 1. P. Artal, A. Guirao, E. Berrio, and D. R. Williams, J. Vision 1, 1 (2001). 2. P. W. de Waard, J. K. Ijspeert, T. J. van den Berg, and P. T. de Jong, Invest. Ophthalmol. Visual Sci. 33, 618 (1992). 3. M. L. Hennelly, J. L. Barbur, D. F. Edgar, and E. G. Woodward, Optom. Vision Sci. 18, 197 (1998). 4. S. W. Chang, A. Benson, and D. T. Azar, J. Cataract Refractive Surg. 24, 1064 (1998). 5. J. A. del Val, S. Barrero, B. Yánez, J. Merayo, J. A. Aparicio, V. R. González, J. C. Postorm, and S. Mar, Appl. Opt. 40, 1727 (2001). 6. J. Buhren and T. Kohnen, J. Cataract Refractive Surg. 29, 2007 (2003). 7. L. Franssen, J. E. Coppens, and T. J. T. P. van den Berg, Invest. Ophthalmol. Visual Sci. 47, 768 (2006). 8. G. Westheimer and J. Liang, J. Opt. Soc. Am. A 12, 1417 (1995). 9. T. Kuroda, T. Fujikado, S. Ninomiya, N. Maeda, Y. Hirohara, and T. Mihashi, J. Refract. Surg. 18, S598 (2002). 10. J. M. Bueno, Vision Res. 41, 2687 (2001). 11. J. M. Bueno, E. Berrio, M. Ozolinsh, and P. Artal, J. Opt. Soc. Am. A 21, 1316 (2004). 12. F. Díaz-Doutón, A. Benito, J. Pujol, M. Arjona, J. L. Güell, and P. Artal, Invest. Ophthalmol. Visual Sci. 47, 1710 (2006). 13. J. Tabernero, A. Benito, V. Nourrit, and P. Artal, Opt. Express 14, 10945 (2006). 14. T. Möller-Pedersen, H. F. Li, W. M. Petroll, H. D. Cavanagh, and J. V. Jester, Invest. Ophthalmol. Visual Sci. 39, 487 (1998). 15. T. Möller-Pedersen, H. D. Cavanagh, W. Matthew Petroll, and J. V. Jester, Ophthalmology 107, 1235 (2000). 16. L. T. Chylack, Jr., J. K. Wolfe, D. M. Singer, M. C. Leske, M. A. Bullimore, I. L. Bailey, J. Friend, D. McCarthy, and S. Y. Wu, Arch. Ophthalmol. (Chicago) 111, 831 (1993). 17. E. Alcon, A. Benito, G. M. Perez, A. De Casas, S. Abenza, S. Luque, J. Pujol, J. M. Marin, and P. Artal, 2007 Annual Meeting Abstract and Program Planner, www.arvo.org, abstract 3822.